Do solar panels damage your roof? The answer depends on something almost nobody checks

A solar quote is a list of decisions: which module, which inverter, what capacity, what mounting structure. All of them are taken now and all of them can be discussed. Underneath that list sits an element that was not chosen, that has no datasheet, and that will carry the system for the next few decades: the roof.

It is the only component of the project that was already there. It arrived with the house, it has an age nobody wrote down, a material chosen by the building custom of the area, and a real condition that is only known once somebody climbs up and looks at it closely. And yet it is what defines what can be installed, how it is fixed and what will happen ten years from now.

Which is why the question that opens this article is nearly always put the wrong way round. It is not whether panels damage the roof, because a module is glass and aluminium and does not even touch the roof covering. It is whether the roof was in a condition to receive them, and whether whoever mounted them took the trouble to check.

And here is the encouraging part: practically everything that can go wrong is decided before the first penetration exists, during a visit that takes one morning. The uncomfortable part is that this visit is exactly what gets trimmed when somebody is in a hurry to close the sale.

A rooftop photovoltaic system is held by anchors, and on most roofs those anchors go through something. That point — the penetration and its seal — concentrates practically all of the project's water risk. Nothing else you see up there has any contact with the inside of your house.

Leaks appear when that point is resolved badly, and the ways of resolving it badly are fairly repetitive. The anchor component is not the one that matches that roof type, so something is improvised out of whatever was in the van. The seal is made with a product that does not survive the weather or the daily heat-and-cool cycle. The penetration lands in the valley of the corrugation or in the overlap, which is exactly where water gathers and runs. Or it is screwed straight onto the sheet, without the sealing component that belongs underneath.

There is a second mechanism, slower and considerably less visible: movement. A metal roof expands and contracts every day, and a structure bolted rigidly onto it works against that movement until a seal gives way. In the same family sits contact between dissimilar metals, which in humid or coastal air accelerates corrosion at precisely the point where you least want it.

None of this is exotic or expensive to avoid. It is the kind of detail an installer with years of roofs behind them solves without thinking, and that an improvised crew discovers in the third rainy season, usually once the phone has stopped being answered.

A properly installed photovoltaic system is measured in decades, and that is precisely its economic virtue: it keeps producing long after the year it finished paying for itself. But that virtue comes with a condition rarely said out loud, because no manufacturer's warranty covers it: the roof holding it up has to reach just as far.

That is where the expense that was in no quote shows up. If the roof needs work halfway through the period — replacing tiles, redoing the waterproofing, repairing a supporting structure that was already compromised — the panels have to come down, be stored, the work done, and everything remounted with its structure, its cabling and its commissioning. It is a complete job, it is paid for separately, and it arrives at the worst possible moment, which is when you had started counting the savings.

The practical conclusion is uncomfortable and very simple: if your roof covering has little life left, the time to redo it is before installing. Doing both jobs back to back costs considerably less than doing one of them twice, and it leaves roof and system starting their lives on the same day.

It is also worth separating age from appearance. A roof can look immaculate from the yard and have its supporting structure eaten by damp; and it can look dirty and faded while being perfectly sound. What settles that is an inspection from above and from below, not a photograph taken from the pavement.

Four or five types of roof coexist on the same Colombian block, and each one imposes a different way of fixing the system. This is why a serious proposal asks about the material before putting a price on the mounting structure: anyone quoting a structure without having asked is quoting for a different house.

On clay tile the job consists of removing the tile at the anchor point, fixing to the batten or to the timber structure, and putting the tile back over it. The classic mistake is resting the fixing on the tile itself, which is not designed to take a point load and ends up cracking. On metal sheet, depending on the profile, there are solutions that clamp to the standing seam without penetrating at all; where a penetration is needed, the fixing goes on the crest of the corrugation and never in the valley.

Fibre cement is the case that demands most caution: it is brittle, it must not be walked on, and older roofs of this material require specific handling precautions that are better resolved beforehand than improvised on site. Sometimes the honest conversation with the client is that this roof should be replaced first.

And on a flat concrete slab the critical point is not the fixing but the water: there are ballasted solutions that do not penetrate the waterproofing, and every one of them has to respect the existing drainage. Blocking an outlet with a badly placed structure is the quickest and silliest way to pond a roof that had worked fine for twenty years.

The roof does not merely hold the system up: it sets its limits. It defines how much usable surface there really is once edges, maintenance walkways and everything already occupying space up there are discounted. It defines the orientation and the tilt the modules will work at. And it defines which shadows will fall on them, and at what times of day.

The irradiation at your location is a fact about the territory and is documented for the whole country. Your production, by contrast, is that fact passed through your particular roof: through its orientation, through the water tank that casts a shadow in mid-afternoon, through the neighbour's party wall, through the aerial, through the tree that is three metres tall today and will be six in five years.

And there is a detail that surprises many people: partial shade can penalise output considerably more than the surface it covers would suggest, depending on how the modules are wired together. Which is why design is not about filling the roof with panels, but about deciding which part of the roof deserves to carry them.

When the roof forces the system to be split across two different orientations — very common in houses with several pitches — that changes the production curve across the day and may also change the electronics worth using. It is a design decision, not a surprise, provided it is taken while looking at the roof rather than at a generic drawing.

The load that most tests a roof carrying panels is usually not the one everybody imagines. The added weight of a photovoltaic system is modest compared with the loads a roof is designed for, although “modest” is not “irrelevant”, and the person who has to confirm it is whoever signs the design — particularly on lightweight roofs or roofs with years behind them.

What really works on the system is the wind. A panel is a flat exposed surface, and wind does not only push: it lifts. That uplift is what defines how many anchor points are needed and how they are distributed, and it is also why a system “securely fixed by eye” may be badly fixed. It is not a matter of tightening more bolts, but of having calculated where they go.

Hail gets talked about a great deal and is worth placing properly: modules are impact tested, and the tempered glass they carry is more resistant than its appearance suggests. In Colombian practice, the useful conversation about hail is not whether the panel survives, but what your policy says if a severe event affects the installation as a whole.

And that is the point almost nobody checks in time. Before installing it is worth calling the building's insurer and asking three things: whether the system is covered, whether it has to be declared, and whether the installation changes any condition of the existing policy. In parallel, ask the installer for their public liability cover. The three questions cost one phone call beforehand and a long argument afterwards.

In a warehouse, a plant or a shopping centre the problem changes scale and an actor appears who is almost never present in housing: the manufacturer of the roof itself. Many industrial metal roofs are handed over with a warranty from that manufacturer, and it can carry very specific conditions about what may and may not be penetrated. Installing without asking risks trading a valid roof warranty for a solar system.

The second actor is the owner of the building, who in many corporate projects is not the party installing. If you operate under a lease, three things have to be in writing before the first rail goes up: who authorises work on the roof, what happens to the system when the lease ends, and who bears the cost of removal if the roof has to be handed back as it was.

Then there is the arithmetic of scale. A large surface means many more anchor points, more drainage to respect and, above all, that a detail done wrong is not done wrong once: it is repeated two hundred times. In a house a badly resolved anchor is a leak; on an industrial roof it is a pattern.

This is why on corporate projects a signed structural calculation stops being a luxury and becomes part of the file, alongside the rest of the technical documentation you are going to need anyway to legalise the system.

None of this requires you to know about roofs. It requires you to ask, and for the answer to arrive in writing, because a proposal that has looked at the roof can answer without discomfort and one that has not shows it on the first question.

The document that most orders the conversation is the site survey report: what type of roof there is, what condition it is in, what anchoring system is proposed and where the planned penetrations go. If nobody went up on the roof before issuing the price, that document cannot exist, and that alone tells you something about what you are comparing.

Then comes the warranty, and here there is a distinction worth money. The equipment warranty is given by the manufacturer and covers the equipment. The watertightness of your roof is covered by nobody by default: RETIE certifies that the electrical installation complies with the technical regulation, which is indispensable, but the waterproofing of the roof is not its subject. It is exactly the gap responsibility falls through when a stain appears on the ceiling and each party points at the other.

And one last detail that looks minor and is not: ask for a photographic record of every sealed point before the panel covers it. It is the only evidence of how the work was done, and the one that cannot be reconstructed afterwards without dismantling half the system.